Very strange bug - 2nd hardrive crash

[OneStiffRod 0]

Here are my comp specs:

2.0ghz INTEL

512MB RAM (rambus)

Nvidia MX 4

Soundblaster Xgamer

Windows 2000 pro

- Here's the prob., I have 2 hardrives the main (C:) is a 10 gig and this is where all of the Flashpoint stuff installed.  The second hardrive is a storage drive 80gig (E:).  

When my computer is using a alot of CPU cycles, IE i am editing/creating a user mission in OFP and thus have websites, folders, and the game open - everything open is in C: drive, the E: 80gig drive is strictly storage and is not being accessed.

When I'm in the OFP Resistance game (in mission editor) everything pauses and I get a message that my 2nd hardrive has been (unplugged/ejected) from the system.  This happens only when I'm using OFP Res and have alot of things open at the same time.  It's happened at least 10 times and mostly I'm able to recover but it is now the 3rd occasion that I am having to REFORMAT the E: drive because incomplete data has been put there by some program causing all data on the strorage (80gig) drive to be lost.  This has nothing to do with the drive itself as it works flawlessly on other PC's in my home and recovers every time this happens but only a few times a complete crash (data loss) has happened as NOW. The drive is connected via EIDE (internal).

It is my opinion that OFP is trying to access the E: drive(storage) when the computer is under stress with multiple programs running.  It then attempts to write data to the E: drive and for some reason it fails and screws everything up by putting incomplete or invalid data on the wrong drive.

I have no viruses and it's too much of a coincidence that it's been OFP running EVERY time this has happened.  

I'm not upset by this but wish if you could look into the way OFP reacts when maybe a buffer is OVERRUN or MEMORY or something else is lack or running behind the game.  I don't see how it could be anything else but an OFP Res. problem.

-Everything I have open are Folders from the C: drive and websites, plus OFP Res when this happens.

Kevin D.

[email protected]

[bn880 5]

To track what's reading what get a program called FileMonitor from Sysinternals.

This can be a starting point. I wouldn't expect OFPR to read from E unless you are starting the game. Also, maybe windows is using the 80 gig for swap space? A friend of mine had an 80 gig maxtor and also lost all data.

[OneStiffRod 0]

Thanks for the info, it is a MAXTOR 80gig.  

I don't know if you know of a program that I could get that would allow me to erase the offending data and restore the drive?  I hate that it only takes like a KB of improper data to mess up Gigabytes worth of data.  

If windows was using the 80gig as a SWAP I still don't get why only OFP resistance would cuase this problem - the older doesn't???

[bn880 5]

OFPR uses more memory than OFP, especially when u are working with Nogova.

Normally improper data can be truncated, not affecting the rest of the drive. In the case of the maxtors, I think for some reason the improper data is in the boot or partition sectors. I really don't know for sure what's causing this. Your motherboard, is it one of the newest ASUS boards?

I'll tell you, you have a different OS then my friend did, also he had AMD.

[OneStiffRod 0]

I have an INTEL motherboard - compatible/made for the INTEL 2.0 ghz chip.

I agree that OFP-R uses more mem but that should deal with graphics/textures and isn't all the graphic elements processed by my Video Card?? which is a near top of the line

NVIDIA GeForce 4 MX w/ 128MB RAM

Maybe it's a problem with the shear size of the MAXTOR drive (80gigs) unpartitioned. Was your friends MAXTOR drive partitioned or like mine in a single clump??

OFP-R is causing it but maybe it's just that it uses so much friggin mem that it triggers the SWAP.

I'd still like to hear from BI on how overloads are handled in the game, such as with buffers and such because maybe it's a mem leak or something.

With 512MB RAM I shouldn't be running outta mem with 3 Folders open and 4 Websites open. It all adds up to no more than 120MB RAM used at most.

[OneStiffRod 0]

YEAHHHHHHHHHHHHH!!!!!!!!!!!!!!

It fixed itself, I had to restart my PC 3 times but I finally got back all the data I lost on the E: drive.

I'm still paranoid that this will happen again as it has before, I just dodged a bullet not having to reformat.

All of my questions still stand though... thanks to anyone who answers.

[Mister Frag 0]

If OFP tried to access memory outside its own address space, an ACCESS VIOLATION exception would be generated, there really is no way that it could randomly cause corruption on a drive that it isn't accessing.

Regarding a possible source of the corruption, make sure that your power supply is up to the task. Lots of power-hungry Rambus memory, a high-speed CPU, 128MB memory card, and multiple drives will put quite a drain on the PSU. I would get a program such as Motherboard Monitor or HMonitor to keep an eye on the different voltages, although you probably won't be able to catch a brief drop in current.

Another possibility is that the hard drive itself has problems and is about to fail. If it supports S.M.A.R.T., enable drive monitoring in the BIOS if you can.

Finally, make sure you format the drive using NTFS, which is far more tolerant of problems than FAT32. It is a journaling file system, which greatly increases the chances of surviving if the system crashes or loses power.

[bn880 5]

Yes he made a nnice 80 gig partition. Even though I told him not to.

Yea you can monitor the board, also try to partition to 20-30 gig just in case.

[OneStiffRod 0]

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote </td></tr><tr><td id="QUOTE">

I've already posted this on another thread, but I think it might give a good impression of the influence of RAM to performance with the ground detail setting:

Specs: AMD TBird 900, 512MB SD133, GF2GTS 32MB Det. 29.80, 32GB 7200rpm HDD, SBLive5.1Player, W2k

Oprah settings:

disabled some shadow features

reduced the cockpit texture setting (don't know how much performance this might bring, but I don't care too much about the cockpit, so what  )

reduced the number of objects(256 is default, I changed to 200)

Light sources down to 20 instead 25 or 26

Viewdistance at ~1000

ground details high

1024x768x32 @75 or 85Hz on 19" monitor

don't know what else

I played a little more around, someone else should verify this results of memory usage:

Main menu with the according map in background (loading a mission on it, but cancel it)

Nogova:

very low: 180MB

high: 248MB

very high: 403MB

Desert Island:

very low: 47MB

high: 120MB

very high: 300MB

Everon:

very low: 125MB

high: 180MB

very high: 384MB

I also noticed that when you reduce the setting the memory is not set free. Smells like a potential issue...

I checked the memory usage of the Resistance process in the TaskManager by ALT-TABbing, so it might not be 100% accurate. (eg. 155% increase from low to high at Desert Island, but only 38%/44% on the others)

If these values are correct, it will be pretty much impossible to play Resistance on a 128MB maschine.

<span id='postcolor'>

This dude said the above, under the Tuning options Thread about midway down in responses. (1st Thread in this Troubleshoot forum)  It points out a memory issue especially with Nogova.

I beleive it's a mem issue - too much being used by OFP-R causes a hickup by windows which then tries to do something it shouldn't and force feed my E: drive which then pukes.

[Mister Frag 0]

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote (OneStiffRod @ Sep. 19 2002,19:28)</td></tr><tr><td id="QUOTE"><snip>

I beleive it's a mem issue - too much being used by OFP-R causes a hickup by windows which then tries to do something it shouldn't and force feed my E: drive which then pukes.<span id='postcolor'>

Speaking as a software/hardware engineer, that theory makes no sense.

[OneStiffRod 0]

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote </td></tr><tr><td id="QUOTE">

Speaking as a software/hardware engineer, that theory makes no sense.<span id='postcolor'>

Let me make of sense of it for you then.

The theory suggested by bn880 in previous posts was that windows was using the E:(storage) drive as a SWAP space.  Not sure what that is exactly, however I know windows uses Virtual MeM and perhaps it may be using the E: drive as such. OR maybe it's that windows has some other problem that when there is too little RAM mem it will cause a failure. Obviously there is data being written to the E: drive that is corrupted when it is unplugged abruptly as I wouldn't get a message from windows that the E: drive is NOW unformatted and needs to be formatted if the case was otherwise.

I don't beleive that the problem is or has anything to do with POWER supply if that were the case my whole system would crash and restart itself.

The problem only appears when I have other TASKS & PROGRAMS (websites) running at the same time I have OFP-R running. The post I just quoted shows that when NOGOVA island is running on high detail...

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote </td></tr><tr><td id="QUOTE">

Nogova:

very low: 180MB

high: 248MB

very high: 403MB

<span id='postcolor'>

... it will use nearly 400MB of RAM.  This SMELLS like I've been running outta RAM mem and the strange effects of this is that my E: drive gets disconeccted and sometimes that disconnect causes complete hardrive crash.

Thanks to the help I've recieved it' s probably safe to assume that OFP-R is not directly responsible for my E: drive probs. but it does appear to have an indirect affect on these problems by using up all of the available RAM which may very well be a mem leak problem as alluded to in other posts in this forum.

The previous post makes complete sense to me and anyone who didn't understand should read it again. All it says is that WINDOWS is causing the CRASH but only because OFP-R is using up all the MEM.

[bn880 5]

Yeah, I think maybe Mister Frag just didn't understand your funny description of it.

He is very knowledgeable in PC hardware and OSs'...

I am not very sure, but it looks like a HD problem at first glance, and ram or CPU problem at second glance. I don't think I can pinpoint it. (could even be the MainBoard HD controller)

[bn880 5]

Maybe a good thing for you to do would be to record the exact errors and problems you get on the E drive, and contact Maxtor somehow? Maybe they are aware of a conflict.?

[Mister Frag 0]

OneStiffRod, my message quoted the exact part of the theory that I had a problem with.

Anyway, it is easy enough to ascertain whether Windows is using the E: drive for swap space -- go into the System applet in the Control Panel and check. If E: is being used and you don't like it, set the maximum size to 0, and possibly increase the swap space on C: to compensate for the reduction.